Sync Docs logo

Sync Docs

Community
onevcat
sync-docs

Keep the agent-facing manual under docs/ in sync with the implementation. Diff-driven from a committed commit baseline; makes conservative, minimal edits only where the code has actually changed user-facing behavior. Run on demand or as part of release prep.

Overview

Publisheronevcat
RepositoryProwl
Skill namesync-docs
Stars
647
Forks
66
Bundled files
Instructions only
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by onevcat on GitHub. Read the source before you install it.

Installation

Install the Sync Docs AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/onevcat/Prowl/tree/main/.claude/skills/sync-docs
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/onevcat/Prowl.git /tmp/Prowl
mkdir -p .claude/skills
cp -r /tmp/Prowl/.claude/skills/sync-docs .claude/skills/sync-docs
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Sync Docs in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Sync Docs on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Sync Docs is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Sync Docs

Keep the agent-facing documentation under docs/ accurate as the implementation evolves. This skill is diff-driven (it only looks at source that changed since a committed baseline) and deliberately conservative (it makes the smallest edits needed to restore accuracy — it does not rewrite, restyle, or expand the docs).

Read docs/README.md first for the doc set's structure and intent. The docs are written for AI agents to read, are plain Markdown, and are designed to be verifiable: each file lists Keywords/Related, and the reference/ files name their source-of-truth.

Guiding principle: be conservative

The goal is to keep facts correct with low churn, not to perfect the prose.

  • Default to leaving a doc unchanged. Only edit when a documented fact is now demonstrably wrong, a documented feature was removed, or a clearly significant new user-facing feature/shortcut/setting was added.
  • Make the smallest possible edit. Fix the wrong cell/line; don't reflow tables, reorder sections, or reword surrounding prose.
  • Tolerate a wide range. Ignore internal refactors, renames, and code reshuffles that don't change user-facing behavior. Ignore stylistic/wording differences. Approximate phrasings ("roughly", "typically", "best-effort") are fine — don't tighten them just because you can.
  • Don't grow the doc set casually. Do not add new sections or new files unless a genuinely new top-level feature appeared. If a change is large (a new feature that needs a whole new component manual, or a structural overhaul), do not do it silently — describe it in the report and let the human decide.
  • Keep "source of truth" pointers correct if files move or are renamed.
  • Prefer touching few files per run. If you find yourself about to edit many files, stop and re-read this section — most of those are probably tolerable.

Steps

  1. Read the baseline. Open docs/.sync-meta.json and read the last_synced_commit field (the last-synced commit hash).

  2. Diff the implementation since the baseline. List source files that changed:

    bash
    git diff --name-only <baseline_commit>..HEAD -- \
      App/Sources/ CLI/ Shared/ ThirdParty/git-wt

    If nothing relevant changed, report "Docs up to date as of <HEAD>." then go to step 6 (bump the baseline) and stop.

  3. Map changed source → docs to re-check. Use this table; only open the docs whose source actually changed. (Most changes touch zero or one doc.)

    Source that changedDocs to re-check
    App/Sources/App/AppShortcuts.swift, App/Sources/Commands/**reference/keyboard-shortcuts.md (+ shortcut mentions in the relevant components/*.md)
    App/Sources/Features/Settings/Models/GlobalSettings.swift, RepositorySettings.swift, UserRepositorySettings.swiftreference/settings-fields.md, components/settings.md, components/custom-actions.md
    CLI/Sources/prowl/**, App/Sources/CLIService/**components/cli.md
    App/Sources/Clients/CLIInstall/**components/cli.md, components/settings.md
    App/Sources/Features/Repositories/**, App/Sources/Domain/Worktree*.swift, App/Sources/Domain/Repository*.swift, App/Sources/Clients/Git/**, ThirdParty/git-wtcomponents/repositories-and-worktrees.md
    App/Sources/Features/Terminal/**, App/Sources/Infrastructure/Ghostty/**components/terminal.md
    App/Sources/Features/Canvas/**components/canvas.md
    App/Sources/Features/Shelf/**components/shelf.md
    App/Sources/Features/CommandPalette/**components/command-palette.md
    App/Sources/Features/ActiveAgents/**components/active-agents.md
    App/Sources/Domain/AgentDetection/**, App/Sources/Infrastructure/AgentDetection/**components/agent-detection.md
    App/Sources/Clients/Notifications/**, WorktreeTerminalState+Notifications.swift, App/Sources/Clients/Dock/**components/notifications.md
    App/Sources/Features/DiffView/**components/diff-view.md
    App/Sources/Clients/Github/**components/github-pull-requests.md
    App/Sources/Clients/Updates/**, App/Sources/Features/Updates/**components/updates.md
    view-mode switching (App/Sources/Features/App/**, ContentView.swift)components/view-modes.md

    overview.md, concepts.md, and README.md only need touch-ups for the addition/removal of a major feature — leave them alone otherwise.

  4. Verify against the source of truth. For each doc identified, check its falsifiable claims against the authoritative source — not against the diff summary — and apply minimal edits per the conservative rules above:

    Claim typeSource of truth
    Keyboard shortcuts (key + modifiers + command ID, remappability)App/Sources/App/AppShortcuts.swift; menu wiring in App/Sources/Commands/*.swift
    Settings field names / types / defaultsApp/Sources/Features/Settings/Models/GlobalSettings.swift, RepositorySettings.swift
    CLI commands / flags / ranges / error codes / JSON fieldsCLI/Sources/prowl/** and App/Sources/CLIService/** (and, if the CLI is installed, prowl <cmd> --help for confirmation)
    Feature behavior / entry pointsthe corresponding App/Sources/Features/** or App/Sources/Clients/**

    When in doubt whether something is a real, user-facing change, leave the doc as-is and flag it in the report rather than editing.

  5. Check internal links if files moved. If you renamed/added/removed any doc file, re-run a quick relative-link check so nothing in docs/ is broken.

  6. Bump the baseline. Set last_synced_commit (and last_synced_date, note) in docs/.sync-meta.json to the current HEAD. This file is committed to git so the baseline persists across sessions and machines — never leave it uncommitted.

  7. Report. Output a short summary:

    ## Docs Sync
    Baseline: <old_hash> → <new_hash>
    Source files changed in range: <count>
    
    ### Updated
    - docs/<file> — <one-line what & why>
    
    ### Checked, left unchanged (tolerated)
    - <area> — <why it didn't need a doc change>
    
    ### Needs human decision (not applied)
    - <large/ambiguous change> — <what & suggested doc action>

Committing

  • Standalone run: stage and commit only docs/** (which includes docs/.sync-meta.json); never git add .. If anything in docs/ changed and you're not on main, open a PR targeting onevcat/Prowl.
  • As part of release prep (the release skill, on main): commit the doc + docs/.sync-meta.json changes as their own commit before the version bump and tag — e.g. git commit -m "Sync docs for <VERSION>". Do not leave them staged/uncommitted: release.sh aborts on a dirty working tree, and the doc commit must already be on main so it becomes an ancestor of the tag and ships inside the release. Bump/tag happen after, never before.
  • Always bump and commit docs/.sync-meta.json even when no doc edits were needed, so the next run starts from a tight diff range. The baseline records the commit the docs were verified against (the current HEAD at run time) — for a release that is the code being shipped, captured before the later doc/bump/tag commits, which is correct (those commits touch no implementation files).

Frequently asked questions

What does the Sync Docs AI skill do?

Keep the agent-facing manual under docs/ in sync with the implementation. Diff-driven from a committed commit baseline; makes conservative, minimal edits only where the code has actually changed user-facing behavior. Run on demand or as part of release prep.

Why use Sync Docs on TypingMind?

Because you install it once and use it with any model. Sync Docs is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Sync Docs in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/onevcat/Prowl/tree/main/.claude/skills/sync-docs. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Sync Docs?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Sync Docs?

As many as you like. As long as a model supports skills, you can use Sync Docs with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Sync Docs AI skill free?

It is published on GitHub by onevcat. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇